Measuring abortion in claims data: what is the state of the science?

Abstract

Health care insurance claims are an increasingly common data source for health outcomes research. While researchers have successfully used several claims data sources for many obstetric and gynecologic questions, use of claims data for abortion and contraception research poses a number of challenges. In this update on the state of the science in identifying abortion in claims data, we review claims data generally, describe commonly used claims data sources, and detail specific reasons why abortion may be underestimated in claims even when employing best practices. We provide examples of successful approaches for identifying abortion in claims, and importantly, spell out limitations when making comparisons across site of care, states, and policy contexts. As increased attention is turned to identifying abortion across diverse settings, it is critical best practices are applied so that the most appropriate inferences regarding abortion incidence across contexts over time are drawn.